基于视差梯度约束的匹配点提纯算法

在图像拼接过程中,粗匹配点对提纯算法RANSAC(Random Sample Consecsus,随机抽样一致性)存在计算量大,效率低等缺陷.为提高该算法提纯匹配点对的效率,基于预检验的策略,提出了基于视差梯度约束的RANSAC算法.实验结果表明,算法在不改变匹配精度的前提下,很好地提高了算法的效率.     

一种基于Matlab的脆弱水印技术

文章分析了现有两个典型的脆弱水印算法的局限性,并充分利用Matlab中rand函数可产生重复随机矩阵的特性以及mean2函数可快速计算矩阵均值的功能,提出了两种基于Matlab的脆弱水印算法。实验结果表明,采用这两种算法实现的水印不仅具有良好的篡改区域定位功能,而且能够有效地克服“拼贴攻击”。     

Sedimentary Micro-phase Automatic Recognition Based on BP Neural Network

IntroductionDuringthedevelopmentofprospectingofthegeologyofpetroleum,itisimportanttoforecastthedistributionofgritstone,mastertheregulationofphysicalparameterinthereservesmasslevel.Especially,itismoreimportanttorecognizetorockphaseandsedimentarycircumstance.Thetraditionalmethodisusuallyqualitativeanalysisbyfetchingrockcoreandreceivingdataaboutrockcoreandrockcrumb.Asthedevelopmentofmeasurementtogeologicwell,theplentifuldataaboutgeologicwellhasbecomeatoolasresearchthesedimentaryfeatures.Combined…